Commentary on passages from Wittgenstein on Rules and Private Language, pp. 38-54

---

• Let no one -- under the influence of too much philosophy of science -- suggest that the hypothesis that I meant plus is to be preferred as the simplest hypothesis. . . . Such an appeal must be based on a misunderstanding of the sceptical problem, or of the role of simplicity considerations, or both. Recall that the sceptical problem was not merely epistemic. The sceptic argues that there is no fact as to what I meant, whether plus or quus. Now simplicity considerations can help us decide between competing hypotheses, but they obviously can never tell us what the competing hypotheses are. If we do not understand what two hypotheses state, what does it mean to say that one is 'more probable' because it is 'simpler'? If the two competing hypotheses are not genuine hypotheses, not assertions of genuine matters of fact, no 'simplicity' considerations will make them so. (K, p. 38).
  
  
  COMMENTARY: This passage, and Kripke's subsequent argument in defense of it, is the subject of a recent paper by me, which appears in The Southern Journal of Philosophy (March 99). Since this paper will eventually be put on this site, I will here highlight only a few of its main complaints against Kripke's case against the appeal to simplicity considerations. To begin with, we should note that it is only in the discussion of simplicity that Kripke contends that there is some sort of problem with understanding what the plus and quus hypotheses are, or what they state. For those, like myself, who not only see no difficulty in understanding the plus and quus hypotheses but also regard it as self-defeating for Kripke's sceptic to suppose that we don't know what the plus and quus hypotheses are or state, can take heart in the fact that no where else in Kripke's book is any doubt cast on our ability to understand the plus and quus hypotheses. We can also take heart in the fact that Kripke's case against appeals to simplicity are at odds with Quine's views. Quine explicitly allows that simplicity considerations not only can be used to choose between competing manuals of translation but allows for the possibility (Quine calls it a remote possibility; I'm not as convinced as he is of its remoteness) that simplicity (and kindred "nonfactual" canons of procedure) allows us to get rid of all but one manual. (See his "Reply to Gibson", in The Philosophy of W.V. Quine.). Quine insists, however, that even in such a case, i.e., simplicity has left us with only one manual of translation, the manual does not give us a fact of the matter about native meaning.
  While it is true that the use of simplicity considerations to scotch, e.g., the quus hypothesis, does not provide a fact of the matter that I or someone else meant/means plus, and so while it is true that simplicity considerations cannot provide a straight solution to the sceptic's challenge, Kripke fails to appreciate that simplicity can provide a sceptical solution to the sceptical problem of showing that someone means plus rather than quus. Following Quine, we can say that the sceptical problem dissolves once we appeal to simplicity, for it is obvious that the plus hypothesis is simpler than the quus hypothesis. As such, the sceptical problem of showing that we meant plus rather than quus is solved, albeit sceptically, i.e., without finding a fact that shows we meant plus rather than quus. Furthermore, a solution based on simplicity considerations has the virtue of blocking KW's case against private language. A sceptical (dis)solution of the sceptic's problem via simplicity considerations is not, a la KW's own solution, inapplicable to a person considered in isolation. Of course, for those benighted enough to want a straight solution to the sceptical problem this result is unsatisfactory. But for those of us, like myself, who willingly grant the nonexistence of facts showing I meant plus rather than quus, the use of simplicity allows for a sceptical solution that doesn't have the impossibility of private language as a corollary.
  One final point here: Kripke's case against the use of simplicity is based, in large part, on confusing, (a) there being no fact(s) that shows I meant plus rather than quus, with (b) there being no fact(s) that show I meant plus or no fact(s) that show I meant quus. Quine warns against a similar sort of confusion. According to Quine, to say that there is no fact of the matter of which of two (or more) competing manuals of translations is the right one is not to be confused with saying that there is no fact of the matter to semantic theories. For Quine, the conformity of a translation manual to speech dispositions is a matter of fact. It's only the choice between competing manuals that lacks factuality. (See "Reply to Roth", in The Philosophy of W.V. Quine ). Similarly, Kripke here supposes that because the hypothesis that I meant plus rather than quus is nonfactual, it follows that the plus and quus hypotheses are also nonfactual. This is simply not the case, as Quine has reminded us. Similarly, the fact that the plus and quus hypotheses do not state genuinely different matters of fact does not warrant saying that the individual hypotheses do not state genuine matters of fact.
  Finally, we should also appreciate that Kripke still writes as if the dispositionalist response has not forced the sceptic to seriously alter the sceptical hypothesis. That is, Kripke fails to appreciate that the dispositionalist has enlarged the set of facts which are available to us. As such, we need to ask which of the two hypotheses, the "we mean plus hypothesis" and the "we mean quus hypothesis", is simpler, given that we are disposed to say, 68 + 57 = 125, and disposed to say, 69 + 57 = 126, etc. In such circumstances, the plus hypothesis is infinitely simpler. In order to hang onto the quus hypothesis here, the sceptic would have to complicate things by also positing a host of factors which would explain why a quusser would be disposed to go on in future cases as a plusser.
  
• Now Wittgenstein's sceptic argues that he knows of no fact about an individual that could constitute his state of meaning plus rather than quus. Against this claim simplicity considerations are irrelevant. Simplicity considerations would have been relevant against a sceptic who argued that the indirectness of our access to the facts of meaning and intention prevents us from ever knowing whether we mean plus or quus. But such merely epistemological scepticism is not in question. The sceptic . . . claims that an omniscient being, with access to all available facts, still would not find any fact that differentiates between the plus and the quus hypotheses. (K, p. 39).
  
  
  COMMENTARY: Kripke is right that simplicity considerations are irrelevant against the sceptic's denial of a fact that shows one means plus rather than quus. But as suggested above, this just means simplicity considerations are irrelevant if one wishes a straight solution to the sceptical challenge. But simplicity considerations can dissolve the sceptical worry that we meant quus in the past, for they justify choosing the plus hypothesis over the quus hypothesis. As such, they are not irrelevant to the sceptical challenge, for by showing how to justify choosing the plus hypothesis over the quus hypothesis, albeit without appeal to facts, they can provide a "sceptical solution" to the sceptical challenge. Indeed, Kripke's language here suggests that he is dimly aware that he has not and cannot make a case against using simplicity considerations to justify choosing between the plus and quus hypotheses. For he here very explicitly tells us that simplicity considerations do not produce, lean on, create, etc., facts of the sort sought by the sceptic. But since KW contends that the best we can do against the sceptic is produce a sceptical solution, the use of simplicity considerations to sceptically solve the challenge, is hardly a compromise solution.
  There is also a problem with the concluding remarks in this passage. Basically, Kripke here argues that although simplicity considerations would have been relevant against epistemological scepticism concerning meaning, KW's sceptic is an ontological sceptic, i.e., someone who calls into doubt the very existence of any facts to show what we mean. The unstated conclusion, of course, is that simplicity considerations are not relevant against the ontological sceptic. Clearly, it's fallacious to argue that because simplicity considerations are, or would have been, relevant against an epistemological sceptic, they are not, or cannot be, relevant against an ontological sceptic. Kripke then is employing an implicit premise to the effect that the only way to beat the ontological sceptic is to produce a meaning fact. Of course, this is both correct but irrelevant. For ultimately KW agrees with his ontological sceptic that there are no meaning facts that show I meant plus rather than quus. But the problem posed by the sceptic is not just that of producing or finding meaning facts. It is the problem of justifying meaning claims. Indeed, this is precisely why the sceptical problem can be sceptically solved, viz., the sceptical problem is not the problem of finding meaning facts (for KW admits this cannot be solved, for there are no such facts) but rather the problem of justifying our meaning claims, our usual meaning talk. Since simplicity considerations can be used to justify meaning claims, they can be used to solve the sceptical problem. So the fact that simplicity considerations cannot produce meaning facts does not show they are irrelevant against an ontological sceptic who claims that the lack of meaning facts undermines our meaning talk.
  Finally, that there is something fishy in Kripke's attempt to dismiss appeals to simplicity considerations is evidenced by the fact that he begins his discussion (the first passage quoted on this page) by suggesting that it doesn't even make sense to speak of one meaning hypothesis being simpler than another, assuming the sceptic is right about the nonexistence of meaning facts. In short, Kripke begins by claiming that simplicity considerations are inapplicable to the meaning hypotheses. But Kripke's final remarks on this matter (the second quoted passage on this page) find him claiming not that simplicity considerations are inapplicable to the meaning hypotheses but rather that they are unable to produce a "meaning fact" and so are impotent to refute the sceptic's denial thereof. Indeed, Kripke's suggestion that simplicity considerations would be relevant against an epistemological sceptic shows that he allows that simplicity considerations can be used to choose between the competing meaning hypotheses. The only "failing" of simplicity considerations, by Kripke's lights, is that a choice based on simplicity does not yield a fact of the matter about what we meant. But as I've noted above, if plus can be said to be the simpler of the two hypotheses (and it can), this gives us all we need for a sceptical solution to the sceptical problem.
  
  
• The idea that we lack 'direct' access to the fact whether we mean plus or quus is bizarre in any case. Do I not know, directly, and with a fair degree of certainty, that I mean plus? Recall that a fact as to what I mean now is supposed to justify my future actions, to make them inevitable if I wish to use words with the same meaning with which I used them before. This was our fundamental requirement on a fact as to what I meant. (K, p. 40).
  
  
  COMMENTARY: This passage goes a long way toward showing what is wrong with the sceptical challenge. Most importantly, if Kripke's presumably rhetorical question, "Do I not know . . . ?", is in fact rhetorical, (and I think it is) then whither the sceptical challenge? If my knowing, directly, with a fair degree of certainty, that I mean plus, is not enough to dismiss the sceptic as a crackpot, what is? At the very least, what it shows is that the sceptical challenge cannot be taken seriously, which in turn shows that somehow, some way, the sceptical challenge is a misguided challenge to our meaning talk. The simple response to the sceptic, and a perfectly natural and appropriate response as well, is to note that, like almost all sceptics, he is requiring certainty when certainty is not required. Furthermore, that we cannot have certainty about our meaning is not a good reason for concluding that our meaning claims are all of them false or nonsensical, which means that we also have not been given a reason to scotch a truth conditional account of meaning in favor of assertion conditions, and no reason then to conclude that private language is impossible. WHEW!
  Another telling claim here is Kripke's telling us what he or KW's sceptic is requiring of a "fact as to what I mean", viz., that it serve to justify future actions, making them inevitable, if I am using words with the same meaning as before. First, Kripke's language betrays him here. No one who holds that meaning is use, (as opposed to assuming abedeutungskorper picture of meaning) would say, "using words with the same meaning as before". Be this as it may, Kripke's fundamental requirement on a "fact as to what I meant" is pretty obviously asking for too much. Given not only that a fact as to what I meant must justify my future action (a thoroughly illegitimate request for anyone who has read Hume on causality and ethics), but also that it must leave no "competing" justifications standing, it should be clear that our alleged inability to find any "meaning facts" is a product of Kripke's perverse requirements on facts. Simply put, if Kripke put the same sort of requirements on "conditions that legitimate the assertion of meaning attributions" (which conditions, according to the sceptical solution, are all we need to legitimate the assertion of meaning conditions; see K, pp. 77-8) as he puts on facts, then it's clear that there is no solution, neither straight nor sceptical, to the sceptical challenge.
  For the assertion conditions for meaning attributions given us by Kripke in chapter 3 neither serve to justify my future responses to mathematical calculations nor rule out competing assertion conditions. As several critics of Kripke have pointed out, whatever basis there is for the community to deem itself plussers, is at the same time a basis for the community to deem itself quussers. There is no way for the community to answer a sceptic who wonders whether the community and its members have not all along been quussing rather than plussing. Without question one of the biggest flaws in Kripke's book is his failure to even address, let alone answer, the worry that the proposed sceptical solution is not asked to answer the sceptical questions posed about meaning facts. By my lights, KW's celebrated sceptical solution comes to little more than "showing" that someone's giving plus answers to computation problems allows us to say, "with a fair degree of certainty", that the person is a plusser, and, supposing that someone gives quus answers to problems where plus and quus require different answers, we're allowed to say, "with a fair degree of certainty", that someone is not a plusser. But since the sceptic puts none of this in doubt, why does Kripke or KW suppose that any of this "solves", sceptically or otherwise, any part of the sceptical problem?
  
  
• Why not argue that "meaning addition by 'plus'" denotes an irreducible experience, with its own special quale, known directly to each of us by introspection? (Headaches, tickles, nausea are examples of inner states with such qualia.) . . . Maybe I appear to be unable to reply just because the experience of meaning addition by 'plus' is as unique and irreducible as that of seeing yellow or feeling a headache, while the sceptic's challenge invites me to look for another fact or experience to which this can be reduced. (K, p. 41).
• I referred to an introspectible experience because, since each of us knows immediately and with fair certainty that he means addition by 'plus', presumably the view in question assumes we know this in the same way we know that we have headaches -- by attending to the 'qualitative' character of our own experiences. (K, p. 41).
• Well, suppose I do in fact feel a certain headache with a very special quality whenever I think of the '+' sign. How on earth would this headache help me figure out whether I ought to answer '125' or '5' when asked about '68 + 57'? If I think the headache indicates that I ought to say '125', would there be anything about it to refute a sceptic's contention that, on the contrary, it indicates that I should say '5'? (K, p. 42).
• No internal impression, with a quale, could possibly tell me in itself how it is to be applied in future cases. Nor can any pile up of such impressions, thought of as rules for interpreting rules, do the job. The answer to the sceptic's problem, "What tells me how I am to apply a given rule in a new case?", must come from something outside any images or 'qualitative' mental states. . . . (Obviously I do not have an image of the infinite table of the 'plus' function in my mind. Some such image would be the only candidate that even has surface plausibility as a device for telling me how to apply 'plus'.) (K, p. 43).
• Of course the falsity of the 'unique introspectible state' view of meaning plus must have been implicit from the start of the problem. If there really were an introspectible state, like a headache, of meaning addition by 'plus' (and if it really could have the justificatory role such a state ought to have), it would have stared one in the face and would have robbed the sceptic's challenge of any appeal. But given the force of this challenge, the need philosophers have felt to posit such a state and the loss we incur when we are robbed of it should be apparent. Perhaps we may try to recoup, by arguing that meaning addition by 'plus' is a state even more sui generis than we have argued before. . . . Such a move may in a sense be irrefutable, and if it is taken in an appropriate way Wittgenstein may even accept it. But it seems desperate: it leaves the nature of this postulated primitive state -- the state of 'meaning addition by "plus"' -- completely mysterious. (K, p. 51).
  
  
  COMMENTARY: These passages bring up a problem that Kripke says much about but which ultimately, the sceptic must admit, does not have a solution. (As the last passage here shows). The problem is the possibility that meaning plus is a unique, sui generis kind of thing, and so cannot be reduced to any non-semantic fact(s) or experience. In short, the sceptic is being accused of assuming something that very well may not be true, viz., meaning is reducible to non-semantic properties. Colin McGinn is Kripke's severest critic in this regard, telling us: "The sceptic is assuming that unless semantic facts can be captured in non-semantic terms they are not really facts; but why should this assumption be thought compulsory? . . . Unless this question can be answered Kripke's sceptic is wide open to the objection that he is mistaking irreducibility for non-factuality: he finds that he cannot provide a non-semantic fact to constitute a semantic fact and then concludes that there are no semantic facts, when the correct conclusion ought to be that semantic facts cannot be reduced to non-semantic facts. (Wittgenstein on Meaning, p. 151).
  Since I am no fan of saying that meaning plus has its own unique quale, nor a fan of saying that meaning plus is any kind of inner state, I am sympathetic to Kripke's conclusion that such appeals are not the way to answer the sceptical challenge. However, a few comments are in order. First, note that whereas the second passage of the set of passages above allows that "each of us knows immediately and with fair certainty that he means addition by 'plus'", the last passage of the set above claims that the force of the sceptical challenge is incompatible with such "immediate knowledge". Odd, to say the least. If we have the immediate knowledge then we know the sceptical argument cannot be right.
  Also, I've often found it interesting that sceptic's never hesitate to charge their opponents with "desperate responses" to their challenges when the most desperate part of the whole story is usually the sceptic's original contention! Surely the sceptic's contention that we meant quus in the past is a paradigm example of a "desperate position". If one is dubious of this, consider the reactions of bank tellers when you suggest to them that you deserve more money in your account because you mean a nonstandard function by 'plus'!
  Finally, these passages allow us to once again see the illegitimate character of the sceptical challenge. The third passage above finds Kripke asking how a headache-like experience could "help me figure out whether I ought to answer '125' or '5' when asked about '68 + 57'?". But this is surely not the problem, viz., that I don't know what I ought to answer. I know I ought to say "125", if I mean plus. I know I ought to say "5", if I mean quus. Presumably, the problem is knowing which function I mean. But this is hardly a problem, for we are the masters, not the words! I tell you which one I mean by telling you that I ought to answer '125' or '5' as the case may be. We could here, as Wittgenstein suggests, speak of a "decision" to mean one thing rather than another but even this is misleading. We really don't decide anything, we just go on a certain way, unhesitatingly, most of the time. Nothing prevents me from going on another way, nothing forces me to go one way rather than another. If this is the sceptic's point, then it's absolutely well taken. But for all of that, my way of going on is also perfectly justified. Similarly, in the fourth passage, we're told that the the sceptic's problem is, "What tells me how I am to apply a given rule in a new case?" But this is not our problem. We know how to apply plus to "new cases" and we know how to apply quus to "new cases" . What we don't know, allegedly, is whether we are, or ought to be, applying plus rather than quus. But this alleged piece of ignorance is not ignorance at all. We know the difference between going on as a plusser and going on as a quusser, and that's all we need to know in order to know how to apply either rule in a new case. As such, we're not ignorant of anything.
  
  
• If Hume is right, of course, no past state of my mind can entail that I will give any particular response in the future. But that I mean 125 in the past does not itself entail this; I must remember what I meant, and so on. Nevertheless it remains mysterious exactly how the existence of any finite past state of my mind could entail that, if I wish to accord with it, and remember that state, and do not miscalculate, I must give a determinate answer to an arbitrarily large addition problem. (K, p. 53).
  
  
  COMMENTARY: This is one of those passages in WRPL that lead one to doubt that Kripke spends much time going over the transcripts of his lectures. Either that or it's a passage in which Kripke shows himself to be obtuse. Kripke begins by noting, correctly, that Hume's sceptical doubts concerning the understanding rule out a "past state of my mind" entailing (logically, of course), a "future response" of mine. As such, the case against such an entailment says nothing specific about meaning, it simply casts doubt on a past state of affairs entailing (logically) a future state of affairs. Most importantly, the so-called new scepticism offered by Wittgenstein seems then to be nothing more than a specific case of Humean scepticism concerning the ability of past events to entail future events. Kripke, as if aware of this, suggests that no one is claiming, or ought to claim, that meaning something in the past, if it is a genuine case of meaning, is enough to entail a future response. For along with meaning something, I must also remember my meaning, and wish to accord with it, and not make a mistake, etc. But, according to Kripke, it's "still mysterious" how a past state of my mind, together with my remembering, my wishing and my not making a mistake, etc. could entail a particular future response. Of course it's mysterious how it can, since it cannot, for Humean reasons!! There is no entailing the future by the past. It cannot be done. The key thing here is that we can and ought to agree with Kripke's sceptic that no past state (whether it be a state of meaning or the state of a piece of bread) entails a future state, even if we remember, wish to accord, don't make a mistake, etc. But the agreement is based on Hume's arguments, not those of Kripke's sceptic. Hume tells us that we cannot entail future responses from past meanings, NO MATTER WHAT and if this is all Kripke's sceptic is claiming then his scepticism is hardly new.
  On the other hand, if one wishes to say something about the relationship between past states of meanings and future responses that doesn't involve entailment, e.g., that it is mysterious how a past state of mind can lead me to feel compelled to respond one way rather than another, I don't see the mystery. In particular, there's certainly nothing mysterious about how a computer is programmed to respond with certain answers to certain addition problems. Ultimately, an account of how we come to respond in such cases is not significantly more mysterious than the computer case. We ignore alternative goings-on because we don't see them and we don't see them because in most cases they're simply not there to be seen. And where they are, e.g. the case of quus, we find that plussing is much more effective for counting and figuring than quussing.
  
  
• . . . ultimately the sceptical problem cannot be evaded, and it arises precisely in the question how the existence in my mind of any mental entity or idea can constitute 'grasping' any particular sense [or meaning] rather than another. The idea in my mind is a finite object: can it not be interpreted as determining a quus function, rather than a plus function? . . . For Wittgenstein, Platonism is largely an unhelpful evasion of the problem of how our finite minds can give rules that are supposed to apply to an infinity of cases. (K, p. 54).
  
  
  COMMENTARY: The sceptical problem we are told is, "How any mental entity can constitute grasping any particular meaning rather than another?" The problem allegedly arises because the idea is finite and so can be interpreted in various ways. But I simply do not see why the possibility of alternative interpretations threatens my understanding of my grasp of a rule or meaning. It's like supposing that because the highway I'm traveling on is dual listed as Highway 70 and Highway 169 that I don't know where I've been or where I'm going, or what I am going to do when I get to the point where the two highways diverge. Clearly this is silly. Most importantly, I do not have to answer the question of which highway I've been traveling on (clearly I've been traveling "both" highways) in order to decide which way to go at the point at which they diverge. That is, my inability to distinguish the "two" highways while traveling on "both" of them has no bearing, no importance, for a decision about where to go at the point of divergence. The sceptical problem then, is an unhelpful invasion into questions concerning our understanding of meanings and rules. That I can find nothing in my past or in my head or anywhere else to distinguish following plus from following quus does not, contra Kripke's sceptic, prevent me from knowing the difference between plussing and quussing, nor does it prevent me from having a justification for going on one way rather than another. It's because I know what quus and plus require in the future that permits me to justify going on one way (e.g., 125) rather than another (e.g., 5). My justification for the former answer rather than the latter is that I wish to plus rather than quus. If you want a justification for my wishing, it's going to appeal to the use I wish to make of the calculation. (N.B. Quussing is as much calculation as plussing). Did I mean plus in the past? Of course. Did I also mean quus in the past? I see nothing to rule out a YES answer but also no reason to be disturbed by a YES answer. It is here that I differ from Kripke's sceptic.

Top of Page | WRPL Ultimate HomePage